Mencoder with multiple TS audio streams

[lanmat]

I've been playing with using mencoder to do the double whammy of excising commercials (with aid of comskip) and to transcode to mpeg4 so my 1.5TB hdd isn't full after a month of HD recording. I think I've finally got a recipe that results in acceptable quality in a file small enough to stream over wireless to my laptop (and using a video codec that doesn't push my cpu usage through the roof, <cough!>x264<cough!>). Mencoder code at bottom of post...comments welcome, hopefully with an answer to my main question:

On almost all programs recorded from my HDHomerun, there are 2 audio streams (I can see them with media player classic under the filters submenu). The second stream is sometimes another language like spanish, but sometimes it's in english and includes a drab narration voice-over. I assume since it pops up as english, it gets included in my transcoded .avi's. I figured out I can use -aid to select the 'non-narration' audio stream, but the actual audio ID changes from program to program. -aid 84 might work for one program to isolate the correct stream, but another requires -aid 52.

Anyone know of a better way to use ONLY the first audio stream or to figure out how to determine which aid is the main audio channel w/o the "extras"?


[SIZE="1"]Current mencoder batch file:
REM Pass #1
"c:\program files\gbpvr\mencoder" %1 -edl "%~dpn1.edlp" -o "%~dpn1.avi" -aid 52 -oac mp3lame -lameopts fast:preset=standard -vf pp=de/md -ovc lavc -lavcopts vcodec=mpeg4:mbd=1:qpel:trell:v4mv:lumi_mask=0.2:vmax_b_frames=1:cmp=3ubcmp=3:autoaspect:vbitrate=4000:vpass=1:turbo:threads=1
REM Pass #2
"c:\program files\gbpvr\mencoder" %1 -edl "%~dpn1.edlp" -o "%~dpn1.avi" -aid 52 -oac mp3lame -lameopts fast:preset=standard -vf pp=de/md -ovc lavc -lavcopts vcodec=mpeg4:mbd=2:qpel:trell:v4mv:dia=-2:lumi_mask=0.2:vmax_b_frames=1:cmp=3ubcmp=3:autoaspect:vqmax=10:vbitrate=4000:vpass=2:threads=1 [/SIZE]

[pBS]

might try "-alang en" to select language, sometimes only gets first english track so may work in your case...at least you know you'll only get English tracks..
typically the first one is the main one of any language..

[lanmat]

Negatory. I suppose since the narration is in english, it gets chosen as well. Here's a snippet of the output when using -v mode for two programs w/ different aid streams:

Program 1:
[SIZE="1"]COLLECT_SECTION, start: 64, size: 184, collected: 184
SKIP: 0+1, TID: 0, TLEN: 13, COLLECTED: 184
PARSE_PAT: section_len: 13, section 0/0
PROG: 5 (1-th of 1), PMT: 80
COLLECT_SECTION, start: 64, size: 184, collected: 184
SKIP: 0+1, TID: 2, TLEN: 55, COLLECTED: 184
FILL_PMT(prog=5), PMT_len: 184, IS_START: 64, TS_PID: 80, SIZE=184, M=0, ES_CNT=
3, IDX=0, PMT_PTR=01573730
...descr id: 0x86, len=7
PARSE_PMT(5 INDEX 0), STREAM: 0, FOUND pid=0x51 (81), type=0x10000002, ES_DESCR_
LENGTH: 9, bytes left: 28
...descr id: 0x81, len=7
PARSE_PMT(5 INDEX 1), STREAM: 1, FOUND pid=0x54 (84), type=0x2000, ES_DESCR_LENG
TH: 9, bytes left: 14
...descr id: 0x81, len=7
PARSE_PMT(5 INDEX 2), STREAM: 2, FOUND pid=0x55 (85), type=0x2000, ES_DESCR_LENG
TH: 9, bytes left: 0[/SIZE]

Program 2:[SIZE="1"]
COLLECT_SECTION, start: 64, size: 184, collected: 184
SKIP: 0+1, TID: 0, TLEN: 13, COLLECTED: 184
PARSE_PAT: section_len: 13, section 0/0
PROG: 3 (1-th of 1), PMT: 48
COLLECT_SECTION, start: 64, size: 184, collected: 184
SKIP: 0+1, TID: 2, TLEN: 55, COLLECTED: 184
FILL_PMT(prog=3), PMT_len: 184, IS_START: 64, TS_PID: 48, SIZE=184, M=0, ES_CNT=
3, IDX=0, PMT_PTR=003FFFC0
PROG DESCR, TAG=81, LEN=7(7)
...descr id: 0x86, len=7
PARSE_PMT(3 INDEX 0), STREAM: 0, FOUND pid=0x31 (49), type=0x10000002, ES_DESCR_
LENGTH: 9, bytes left: 19
...descr id: 0x81, len=7
PARSE_PMT(3 INDEX 1), STREAM: 1, FOUND pid=0x34 (52), type=0x2000, ES_DESCR_LENG
TH: 9, bytes left: 5
PARSE_PMT(3 INDEX 2), STREAM: 2, FOUND pid=0x35 (53), type=0x2000, ES_DESCR_LENG
TH: 0, bytes left: 0[/SIZE]